README.md

PatentsLens.jl

Julia package for handling Lens.org patent data.

Installation

Adding the registry

All packages in the JuliaPatents family are registered in the JuliaPatents registry. To add the registry, enter the julia REPL and run:

using Pkg
pkg"registry add https://github.com/JuliaPatents/Registry"

This only needs to be done once.

Adding the package

After adding the registry, the package can be added to any Julia environment:

using Pkg
pkg"add PatentsLens"

The package can now be loaded:

using PatentsBase, PatentsLens

Optional analysis interface

PatentsLens.jl implements the analysis interface defined by PatentsLandscapes.jl. To use it, import the PatentsLandscapes package (included in this package's dependencies):

using PatentsLandscapes

Getting started

The main purpose of this package is to import patent metadata as exported from Lens.org in the jsonlines (.jsonl) format.

The package supports two data models:

• An in-memory object model similar to the original JSON, using Julia structs
• An SQLite-based relational model that offers indexed and fast property-based and full-text search, aggregation, and more

The PatentsLandscapes.jl API is currently only implemented for the SQLite model.

Using the object model (in-memory)

Loading data from a file test.jsonl into memory looks like this:

applications = PatentsLens.read_jsonl("test.jsonl")

The LensApplication struct implements the interface defined in PatentsBase.jl.

The dataset can easily be elevated to the simple family level:

families = PatentsLens.aggregate_families(applications)

Using the SQLite database

To begin using the SQLite model, create a new database:

db = LensDB("database.db")

This will create a new SQLite database at the specified path and initialize it with the PatentsLens schema.

Data can then be loaded into the new database like this:

PatentsLens.load_jsonl!(db, "test.jsonl")

Because the data needs to be transformed into a relational form, this step may take a while.

Data can be retrieved from the database and converted back into the object model:

all_apps = applications(db)
all_fams = families(db)

Subsets of the data can be created and accessed using PatentsBase's Filter API:

# Retrieve all patent families from the database that mention polylactic acid in their abstract
pla_fams = families(db, ContentFilter("polylactic OR PLA", AbstractSearch()))